METHODS

Study population

This study was approved by the ethics committee of Singapore National Eye Centre and carried out in accordance with the tenets of the World Medical Association’s declaration of Helsinki. Informed, written consent was obtained from all subjects. The study was part of a population based survey of ocular disorders among adult Chinese living in Singapore, conducted between 10 October 1997 and 14 August 1998. Detailed population selection and methodology have been previously reported.^40–42 In brief, the 1996 Singapore electoral register in the district of Tanjong Pagar was used as the sampling frame in this study. Tanjong Pagar was chosen because the population demographics of the Chinese residents were representative of the rest of Singapore. The electoral register listed 15 082 Chinese names of people aged 40–79 years residing in the district. Two thousand (13.3%) names were initially selected using a disproportionate (with more weights given to the older age groups), stratified, clustered, random sampling method.

Procedures

The grading of lens opacity was performed by trained ophthalmologists (90% by PJF), according to a written standardised protocol, using the Lens Opacity Classification System (LOCS) III system.⁴³ After dilatation of pupils with tropicamide 1% and phenylephrine hydrochloride 2.5% eye drops (repeated twice if necessary), the participant was examined at a slit lamp (Model BQ 900, Haag-Streit, Bern, Switzerland), and the presence and severity of specific lens opacity were compared and documented according to LOCS III standard photographs.

Trained study personnel, masked to cataract status, gathered information on demographic, education, occupation, income, medical history, as well as other variables from a standardised interview. In addition, height (m) was measured without shoes. Weight (kg) without shoes was measured on a single automatic weighing scale. Body mass index (BMI) was derived from the ratio of person’s weight divided by the square of his height (US National Institutes of Health cites normal range 18.5–24.9). Systolic and diastolic blood pressures were measured (mm Hg).

For non-responders to the clinic examination an abbreviated examination was conducted in their homes. Cataract was similarly graded using LOCS III standards with a portable slit lamp (Model 904, Clement Clarke, Harlow, Essex, UK), after dilatation of the pupils. Interview data were also similar, but height, weight, and blood pressure measurements were not available for most of these people.

Definitions of cataract

The LOCS III classification system was used for grading lens opacity.⁴³ This is a method of grading severity of lens opacities according to photographic standards, separated into four major groups: nuclear opalescence (NO), nuclear colour (NC), cortical (C), and posterior subcapsular (P). Because there is a wide spectrum of lens changes and different types of opacities may be present in one or both eyes, three definitions were used for this analysis.

Nuclear, cortical, and posterior subcapsular cataracts

This definition included the presence, in at least one eye, of significant nuclear (NO and NC), cortical, and posterior subcapsular lens cataract. Significant nuclear cataract was the presence, in at least one eye, of a LOCS III score of ⩾4 for NO or ⩾4 for NC. Significant cortical cataract was the presence, in at least one eye, of a LOCS III score of ⩾2 for C. Significant posterior subcapsular cataract (PSC) was the presence, in at least one eye, of a LOCS III score of ⩾2 for P. If a person had unilateral cataract surgery or a non-gradable lens, the LOCS III score of the fellow eye was used. Definitions of cataracts were based on similar criteria published by other groups using the LOCS III system.^9,44

Any cataract

This was defined as the presence, in at least one eye, of significant nuclear, cortical, or posterior subcapsular cataract, as defined above (that is, LOCS III score of ⩾4 for NO or ⩾4 for NC or ⩾2.0 for C or ⩾2.0 for P).

Any cataract surgery

This was defined as the presence, in at least one eye, of a history of previous cataract surgery (aphakia or pseudophakia).

Definitions of other variables

Age was the age at time of examination. Education was ascertained by the question, “What was your highest education level?” and categorised into four groups: no formal education, primary (6 years or less), secondary (7–10 years), and tertiary (11 years or older, including university education). Occupation was ascertained with the question, “What group of occupations do you feel best categorises your job?” with the response categorised into six groups for this analysis: managerial and professionals; clerical workers and sales people; production workers and machine operators; labourers and agricultural workers; homemakers, and others. Housing type was categorised into three groups for analysis: one or two room government flats, three room government flats, and government flats with four or more rooms; executive government flats; or private housing. Individual and household monthly income were recorded (in Singapore dollars, current exchange rate of 1 = Sing$2.80, $1 = Sing$1.77), with individual income categorised as: S$1000 or less, S$1001–2000, S$2001–3000, and more than S$3000 (retired people were excluded from individual income analysis, n = 119). Household income was categorised as less than S$2000, S$2001–4000, and more than S$4000. Medical history of diabetes, hypertension, heart attack, and stroke were ascertained by asking, “Have you been told by a doctor that you have diabetes (hypertension, heart attack, stroke)?” followed by further questions on treatment and hospitalisation. Hypertension was defined as a self report of hypertension or a systolic blood pressure of ⩾160 mm Hg and/or diastolic blood pressure of ⩾95 mm Hg, if measurements were available. For analysis, hypertension, heart attack, and stroke were categorised as present or absent. For diabetes, a separate classification included diabetes on treatment with oral diabetic medications and/or insulin injection, diabetes on diet treatment only, and no diabetes. A history of current cigarette smoking was ascertained by asking, “Do you smoke regularly (at least once a week)?” with a further question on the number of cigarettes smoked per day. Smoking was categorised as smoking more than 10 cigarettes per day, smoking 1–10 cigarettes per day, and non-smoker.

Statistical analysis

The analyses were conducted as follows. The associations between age, sex, socioeconomic variables, medical history, and the presence of specific types of cataract, any cataract, and cataract surgery were evaluated initially by contingency tables and estimated by the odds ratio and its 95% confidence interval (95% CI). Multiple logistic regression models were used to estimate the age and sex adjusted OR, and to assess the influence of the other variables (see Table 2) for each cataract type, any cataract, and cataract surgery. Variables considered for inclusion in the regression models include those that had a p value of <0.10 in the age and sex adjusted models in at least one cataract type. Age and sex were included in all models, in order to adjust for the effect of the disproportionate sampling strategy. Trends in proportions were tested for significance using the Mantel-Haenszel procedure. Statistical analyses of the data were carried out using SPSS (SPSS Inc, Chicago, IL, USA).

RESULTS

Among the 2000 people selected, 46 had died and 235 had moved to addresses outside the district before the study period, and two people were excluded because of severe illness, leaving 1717 subjects considered eligible to participate in this study. These people were invited for a comprehensive eye examination at the study clinic, following which an abbreviated domiciliary examination on non-respondents was conducted. The total number of subjects examined in either setting was 1232 (1090 clinic, 142 at home), an overall response rate of 71.8% (1232/1717). Of these, six people did not undergo pupil dilatation because of narrow angles or angle closure glaucoma, and another 20 were excluded because of bilateral missing lens data (unfit or refused lens examination, or bilateral phthisis bulbi), leaving 1206 people (70.2%) for this analysis.

Of the 1206 people in this study, 103 (8.5%) had a previous cataract operation in the right eye, leaving 1103 potentially gradable lens in the right eye. Likewise, 94 (7.7%) people had cataract operation in their left eyes, leaving 1112 potentially gradable left eyes. The overall number of people with cataract surgery in either eye was 134 (11.1%). There were 621 (54.7%) people with any cataract (nuclear, cortical, or PSC) in either eye. Of these, 450 (40.1%) had nuclear cataract, 431 (38.6%) had cortical cataract, and 130 (12.0%) had PSC cataract (categories not mutually exclusive). Table 1 shows the demographic profile of the study population. Men were more likely to have higher education levels, occupations as managers or professionals, production or machine operators, and higher individual incomes.

Table 2 gives the number people with specific types of cataract, any cataract, and cataract surgery, subdivided according to age, sex, socioeconomic and medical factors. Table 3 summarises the relation between cataract morphology, previous cataract surgery, and various putative risk factors. Increasing age was strongly associated with all types of cataract and cataract surgery (p<0.001 for all categories). After adjusting for age and sex, the following associations were noted. For nuclear cataract, significant associations were found for all non-managerial or non-professional occupations (OR between 2.1 to 2.7), current cigarette smoking, and lower BMI. For cigarette smoking and BMI, there was evidence of a dose-response pattern. People who smoked more than 10 cigarettes per day (OR 2.3) had higher odds than those who smoked 1–10 cigarettes per day (OR 1.5), compared to those who did not smoke (test of trend, p=0.001). People with the lower quintiles of BMI had higher odds of nuclear cataract (test of trend, p=0.03). Finally, a weak association with living in smaller housing units were found (OR 1.5, p=0.08).

Cortical cataract was associated significantly with diabetes and lower BMI. We could not identify an association with cigarette smoking. People with diabetes using oral medication and/or insulin injection have higher odds (OR 3.0) than those on diet control only (OR 1.6), compared to those with no diabetes (test of trend, p <0.001). There was a shallow dose-response relation for BMI, with the lowest quintile having 1.7 greater risk of cortical opacity than the highest BMI quintile (test for trend, p=0.05). For occupational groups, only production workers or machine operators had higher odds compared to managers and professionals. PSC cataract was similarly associated with diabetes, living in smaller housing units, and lower household income.

The presence of any cataract was significantly associated with diabetes, cigarette smoking, lower BMI, and being a production worker or machine operator. Previous cataract surgery was associated with diabetes, and possibly lower BMI.

Multiple logistic regression models were derived for each cataract type, any cataract, and cataract surgery (final models presented in Table 4). Increasing age was a risk factor for all types of cataract, and for cataract surgery. Nuclear cataract was independently associated with cigarette smoking, and non-managerial or non-professional occupation (production or factory worker and labourer or agricultural worker). Cortical cataract was independently associated with diabetes, lower BMI, and non-managerial or non-professional occupation. PSC cataract was associated with diabetes and lower household income. Cataract of any type was associated with diabetes, lower BMI, and non-managerial or non-professional occupation. For cataract surgery, diabetes was the only independent risk factor.

DISCUSSION

Numerous risk factors for the different types of age related cataracts have been identified in white and black populations.^5–10,33,34 These include environmental factors such as ultraviolet or sunlight exposure,^11–15 systemic diseases such as diabetes,^16–21,34 lifestyle factors such as cigarette smoking,^22–27 indices of nutrition (for example lower body mass index),^28,30 and socioeconomic factors such as lower educational attainment.^10,32 Our study indicates similar factors are associated with age related cataracts in Chinese Singaporeans. As in other studies, we found varying patterns and strengths of association for different cataract types, supporting the hypothesis that distinct causal pathways may exist for nuclear, cortical, and PSC cataracts.² Previous studies have identified higher rates of cortical lens opacity in women than in men,^5,9,10,45 although we did not detect any such association.

Studies of risk factors in Asian populations are summarised in Table 5.^2,46–49 In this discussion we have tried to highlight significant differences compared with other populations, and address implications of these findings to the Chinese people.

In European and African derived populations, diabetes has consistently been identified as a risk factor for age related PSC and cortical, but probably not nuclear cataract.^{5,6,8,9,9,10,19–21} In contrast, cigarette smoking is associated with nuclear and possibly PSC, but not cortical cataracts.^{8–10,25–27} In a population based study of Chinese people in Taiwan, diabetes was associated with PSC cataract (OR: 1.6, 95% CI: 1.1 to 2.3), while cigarette smoking was associated with nuclear cataract (OR: 1.3, 95% CI: 1.0 to 1.7).³⁹ Our study supports these findings. The consistency of these data in diverse populations therefore suggests these associations are causal in nature. Perhaps more importantly, diabetes and smoking appear to be a growing public health problem in Singapore, Taiwan, and other rapidly developing east Asian countries. Alarmingly, the prevalence of diabetes in Singaporeans aged 18–69 years has risen fourfold within three decades, from 2% in 1975, 4.7% in 1984, 8.6% in 1992, to 9.0% in 1998.^50,51 The increasing prevalence of diabetes reflects increasing affluence, an ageing population, a probable trend towards a more sedentary lifestyle, and dietary changes. Similar epidemiological patterns and trends are seen in Taiwan,⁵² Hong Kong,⁵³ and China,^54–57 suggesting that the risk of cataract attributable to diabetes is likely to increase in the future. A related problem is the escalating prevalence of cigarette smoking observed in Singapore⁵⁸ and China.^59,60 It has been estimated that up to 20% of cataract in a population is related to cigarette smoking.² Thus, an effective antismoking programme in Asia may decrease the burden of cataract blindness, in addition to its potentially beneficial impact on respiratory and cardiovascular health.

The relation between lower BMI and cortical cataract (and possibly nuclear cataract significant after age and sex adjustment, but not after adjustment for other variables) in our study is difficult to explain. There is no overall consensus on either the direction or nature of the relation between BMI and cataract in the literature, partly because it is difficult to infer what BMI represents. A lower BMI has been hypothesised to reflect nutritional deprivation and lower socioeconomic status, particularly in developing countries,⁴⁸ but a higher BMI may also be associated with diabetes, hypertension, and other morbidities.³¹ Thus, while some studies have found lower BMI to be associated with cataract,^10,31,35,36 others have found that a higher BMI increases cataract risk.^28,30 In our population, lower BMI was correlated with cigarette smoking and lower socioeconomic status (education, income, and housing type). It is possible that the associations we observed were related to some of these factors, although the associations were independent of various socioeconomic indicators. We did consider the possibility of a reverse J-shaped relation between cataract and BMI, although more detailed analysis did not support this hypothesis.

In other populations, people of lower socioeconomic status were consistently more likely to have cataract.^{5,8,10,32,33,35,36} In our study, lower education was associated with nuclear cataract, after controlling for smoking, while lower household income was associated with PSC cataract, after controlling for diabetes. Like BMI, the underlying reasons are complex. Lower socioeconomic status and less education are correlated with cigarette smoking, alcohol drinking, lower vitamin supplement intake, and poorer nutrition, all of which may be related to cataract.^61–63 However, other unmeasured factors may also have a role. In the Beaver Dam Eye Study, lower education was associated with nuclear and cortical cataract, while lower income was associated with cortical and PSC cataract, even after controlling for age, sex, diabetes, multivitamin use, alcohol consumption, and cigarette smoking.³² Some of these (for example, multivitamin use, alcohol consumption) may explain or confound the associations in our study. Another possible reason is that people with lower education or income are less likely to see an eye care provider or have cataract surgery performed. However, we did not find any association between socioeconomic indicators and cataract surgery. Indeed, the crude rate of previous cataract surgery in our cohort was high (12% in men, 10% in women). However, our disproportionate sampling strategy makes crude rates misleading. The age corrected rate of 5.1%⁶⁴ is not dissimilar to rates seen in The Blue Mountains Eye Study (6%)⁴⁵ and The Beaver Dam Eye Study (3.6%).⁶⁵ Irrespective of the underlying reasons, public health efforts should be directed at these high risk groups.

Blue collar workers (production or agricultural workers, machine operators, and labourers) were more likely to have nuclear cataract than the managerial and professional classes, after controlling for age, sex, and smoking. Interpreting these findings in a study such as this is difficult. The differences may reflect socioeconomic (probably nutritional) inequities. Environmental factors such as exposure to ultraviolet radiation may account for some of these observations, although we identified no excess risk in outdoor workers (agricultural workers and labourers), compared with production workers in an indoor environment. A study quantifying individual ultraviolet radiation in different occupational groups would be required to satisfactorily address this question.

Few studies have examined associations with cataract surgery as the end point. In our study, diabetes was the only independent risk factor. In Beaver Dam, people with diabetes were also more likely to have cataract surgery within 5 years compared to those without.⁶⁶ In a case-control study in France, diabetes, smoking, and lower socioeconomic status were associated with cataract surgery.⁹

Potential limitations should be noted. Firstly, the reliability of our lens grading is not known. Although we did not have reproducibility experiments (as in the Lens Opacity Case Control Study¹⁰), the grading was conducted according to a written standardised protocol, with over 90% by one ophthalmologist, and there were regular feedback meetings to discuss quality control issues. Further, our findings were comparable to other studies in many other respects for different cataract types. A second possibility was that selection bias significantly affected the results, as our overall participation rate was 70%. For example, the association between lower socioeconomic status and cataract could be explained if people with higher socioeconomic status and cataract were less likely to participate in our study. However, the reverse may be true.

In summary, we found risk factors for the different morphological types of cataract in Chinese people in Singapore were broadly similar to those identified in people of European and African origin, and those from the Indian subcontinent. Diabetes appears to be an important risk factor for cortical and PSC cataract, but not nuclear cataract. Cigarette smoking appears to be associated with nuclear cataract only, while lower BMI is associated with cortical cataract only. Lower socioeconomic status appears to be related to all cataract types. Our findings therefore support the concept that similar mechanisms are important in the pathogenesis of age related cataract, irrespective of ethnic or racial origin.